Upgrade to ww 0.3.1 by freddyaboulton · Pull Request #2181 · alteryx/evalml

freddyaboulton · 2021-04-22T21:19:40Z

Pull Request Description

Fixes #2035
Fixes #2019
Fixes #2285
Fixes #1955

Feature branch tracking the woodwork upgrade. Creating it now so that I can track if we ever fall behind main.

Roadmap:

Demo datasets, utils, and preprocessing
Data Checks
Objectives (this will require updating the confusion matrix impl)
Components (will skip stacked ensembler unit tests since they take in pipelines)
Pipelines (will add stacked ensembler unit tests back in)
automl
model_understanding
Docs

After creating the pull request: in order to pass the release_notes_updated check you will need to update the "Future Release" section of docs/source/release_notes.rst to include this pull request by adding :pr:123.

* Update demos, utils, and preprocessing * Updating make test commands * Update ww requirement * Add test that infer_feature_types preserves schema * Add test that infer_feature_types raises errors with invalid schema * load_data always returns woodwork info. Deleted return_pandas * Updating docstrings

codecov · 2021-04-22T21:22:28Z

Codecov Report

Merging #2181 (cac9870) into main (bcfd02f) will decrease coverage by 0.1%.
The diff coverage is 100.0%.

@@           Coverage Diff            @@
##             main   #2181     +/-   ##
========================================
- Coverage   100.0%   99.9%   -0.0%     
========================================
  Files         280     280             
  Lines       24425   24360     -65     
========================================
- Hits        24402   24333     -69     
- Misses         23      27      +4

Impacted Files	Coverage Δ
evalml/__init__.py	`100.0% <ø> (ø)`
evalml/data_checks/data_check.py	`100.0% <ø> (ø)`
...alml/objectives/binary_classification_objective.py	`100.0% <ø> (ø)`
evalml/objectives/cost_benefit_matrix.py	`100.0% <ø> (ø)`
evalml/objectives/fraud_cost.py	`100.0% <ø> (ø)`
evalml/objectives/lead_scoring.py	`100.0% <ø> (ø)`
evalml/objectives/sensitivity_low_alert.py	`100.0% <ø> (ø)`
evalml/pipelines/components/component_base.py	`100.0% <ø> (ø)`
...tive_tests/test_binary_classification_objective.py	`100.0% <ø> (ø)`
evalml/utils/__init__.py	`100.0% <ø> (ø)`
... and 148 more

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update bcfd02f...cac9870. Read the comment docs.

* Updating data checks * Undo accidental edit * Fixing typo where we assign to ww.init

* Updating objectives * Deleting superfluous pd.Series

gsheni · 2021-04-26T13:41:26Z

Just as a heads up (I think i've brought this up before), in infer_feature_types there is a copy call. If a user passes a 2GB dataset to AutoML, will this take up ~4GB of memory while running AutoML Search?

freddyaboulton · 2021-04-26T18:24:12Z

@gsheni I totally get what you mean!

Our components already copy the data (I believe this is because we have a convention that we should not modify the user's data) so I don't think the copy in infer_feature_types is "new behavior". This change just makes it so that the copy happens in one place, so we can get rid of the copy in our components:

Imputer

LSA

I think you have a point that our convention may cause problems for large datasets. But I don't think that's specific to the ww upgrade. I'd be happy to continue that conversation in a different issue!

* Update components - first commit * Update delayed_features_transformer to not use assign * Fixing tests * Skipping Boolean with Nan test in imputers * Fixing base sampler _prepare_data * Fixing target imputer null bool test * Fix test skips * Addressing comments * Clean up sampler tests * Editing docstrings

* Updating pipelines * Fixing docstrings * Fix stacked test * Add tests to check input data not modified * Use rename in component graph

* Update model understanding module * Removing unused import

freddyaboulton · 2021-05-21T19:08:34Z

          DOCKERHUB_PASSWORD: ${{ secrets.DOCKERHUB_PASSWORD }}
        run: |
-          git clone -b latest_release_changes --single-branch https://github.com/conda-forge/evalml-core-feedstock
+          git clone -b ww-update-branch --single-branch https://github.com/conda-forge/evalml-core-feedstock


I can change this back but the conda check will be red on this pr. Just fyi that might mean that @dsherry needs to merge this in.

@freddyaboulton got it. What do you want to do?

If we're gonna merge it red, please write up an explanation of why its red so others can follow along.

Sorry if I missed this, why can't we add the woodwork upgrade to latest_release_changes?

@ParthivNaresh because the check for all of the other PRs will fail since they haven't merged these changes in and the ww version that would be installed is 0.3.1. But since main has already been frozen for a while, I think the cleanest thing to do now is change the ww version in latest_relase_changes and merge this in!

dsherry · 2021-05-21T21:37:11Z

👀
👀
🤯 😁

dsherry

@freddyaboulton incredible!!!

🎊🚢 🎊

This PR's vibe

bchen1116

This is the wildest PR I've ever seen. 🙌

bchen1116 · 2021-05-24T14:00:25Z

 shap>=0.36.0
 texttable>=1.6.2
-woodwork==0.0.11
+woodwork==0.3.1


nit: can we set to >=0.3.1?

This was restricted to specific Woodwork version because early on there was breaking changes.
We have been more proactive noting the breaking changes in Woodwork in the release notes, and incrementing the version properly.

https://woodwork.alteryx.com/en/latest/release_notes.html#breaking-changes

chukarsten · 2021-05-24T14:19:03Z

👀
👀
🤯 😁

I think it's nuts that the overall net line count is just "-2"

chukarsten

Your Requiem, @freddyaboulton . Bravo.

ParthivNaresh

Absolute legend mate

ParthivNaresh · 2021-05-24T16:03:06Z

          DOCKERHUB_PASSWORD: ${{ secrets.DOCKERHUB_PASSWORD }}
        run: |
-          git clone -b latest_release_changes --single-branch https://github.com/conda-forge/evalml-core-feedstock
+          git clone -b ww-update-branch --single-branch https://github.com/conda-forge/evalml-core-feedstock


Sorry if I missed this, why can't we add the woodwork upgrade to latest_release_changes?

freddyaboulton added 5 commits April 22, 2021 17:22

Release notes for first PR

28346c1

Deleting redundant tests

660210d

update test name

0d2b5ae

WW 0.2.0: Update Data Checks (#2182)

6a8207d

* Updating data checks * Undo accidental edit * Fixing typo where we assign to ww.init

Merge branch 'main' into 2035-use-ww-accessor

725320d

freddyaboulton mentioned this pull request Apr 23, 2021

WW Accessor: Updating objectives #2185

Merged

freddyaboulton added 2 commits April 23, 2021 16:29

WW Accessor: Updating objectives (#2185)

61b8443

* Updating objectives * Deleting superfluous pd.Series

Merge branch 'main' into 2035-use-ww-accessor

edb41ea

freddyaboulton added 2 commits April 26, 2021 11:05

Merge main

617d244

Merge branch 'main' into 2035-use-ww-accessor

40c3a86

freddyaboulton mentioned this pull request Apr 26, 2021

WW Accessor: Updating Components #2191

Merged

freddyaboulton added 2 commits April 29, 2021 10:14

Merge branch 'main' into 2035-use-ww-accessor

b33cb7c

freddyaboulton mentioned this pull request Apr 29, 2021

WW 0.2.0 Update: Updating pipelines #2205

Merged

freddyaboulton added 9 commits May 3, 2021 11:44

Merge pipeline api change

c0e06c9

Merge pct_null_rows

2e73d48

Merge "delete baselines"

d1d97c7

Merge splitter deletion

c4ef949

Update to ww 0.3.0

3987b1f

Fixing components that didn't have merge conflicts

fa2353e

Merge max_depth=1 PR

90578c7

WW 0.2.0 Update: Updating pipelines (#2205)

0e65c58

* Updating pipelines * Fixing docstrings * Fix stacked test * Add tests to check input data not modified * Use rename in component graph

latest dependency changes

e728ec5

freddyaboulton changed the title ~~Upgrade to ww 0.2.0~~ Upgrade to ww 0.3.0 May 10, 2021

freddyaboulton mentioned this pull request May 10, 2021

WW Accessor: Update Automl #2243

Merged

WW Accessor: Update model understanding (#2247)

829912e

* Update model understanding module * Removing unused import

Update core-requirements.txt

6871e8d

freddyaboulton commented May 21, 2021

View reviewed changes

freddyaboulton added 5 commits May 21, 2021 15:19

Delete _convert_woodwork_types_wrapper completely

c4bdd03

Fixing coverage

acc3a5d

Add back standard scaler try

9e386d6

Remove _convert_woodwork_types_wrapper from docs'

af2de08

Linting docs

61a3131

freddyaboulton marked this pull request as ready for review May 21, 2021 21:27

auto-assign bot assigned freddyaboulton May 21, 2021

freddyaboulton requested review from ParthivNaresh, angela97lin, bchen1116, chukarsten, dsherry and jeremyliweishih May 21, 2021 21:27

dsherry approved these changes May 21, 2021

View reviewed changes

bchen1116 approved these changes May 24, 2021

View reviewed changes

chukarsten approved these changes May 24, 2021

View reviewed changes

dsherry mentioned this pull request May 24, 2021

Component graph and text featurizer: preserve user-specified woodwork types #2297

Merged

ParthivNaresh approved these changes May 24, 2021

View reviewed changes

freddyaboulton added 3 commits May 25, 2021 06:11

Merge branch 'main' into 2035-use-ww-accessor

f548c51

Merging main

cdbf713

Fix release notes

cac9870

freddyaboulton merged commit a126c46 into main May 25, 2021

chukarsten mentioned this pull request Jun 2, 2021

v0.25.0 #2330

Merged

freddyaboulton deleted the 2035-use-ww-accessor branch June 3, 2021 14:03

chukarsten mentioned this pull request Jun 23, 2021

Release v0.27.0 #2428

Merged

chukarsten mentioned this pull request Aug 9, 2022

Audit the Unit Test Suite for Woodwork Inference Sensitivity #3651

Closed

Conversation

freddyaboulton commented Apr 22, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Pull Request Description

Uh oh!

codecov bot commented Apr 22, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

gsheni commented Apr 26, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

freddyaboulton commented Apr 26, 2021

Imputer

LSA

Uh oh!

freddyaboulton May 21, 2021

Choose a reason for hiding this comment

Uh oh!

dsherry May 21, 2021

Choose a reason for hiding this comment

Uh oh!

ParthivNaresh May 24, 2021

Choose a reason for hiding this comment

Uh oh!

freddyaboulton May 25, 2021

Choose a reason for hiding this comment

Uh oh!

dsherry commented May 21, 2021

Uh oh!

dsherry left a comment

Choose a reason for hiding this comment

Uh oh!

bchen1116 left a comment

Choose a reason for hiding this comment

Uh oh!

bchen1116 May 24, 2021

Choose a reason for hiding this comment

Uh oh!

gsheni May 24, 2021

Choose a reason for hiding this comment

Uh oh!

chukarsten commented May 24, 2021

Uh oh!

chukarsten left a comment

Choose a reason for hiding this comment

Uh oh!

ParthivNaresh left a comment

Choose a reason for hiding this comment

Uh oh!

ParthivNaresh May 24, 2021

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

freddyaboulton commented Apr 22, 2021 •

edited

Loading

codecov bot commented Apr 22, 2021 •

edited

Loading

gsheni commented Apr 26, 2021 •

edited

Loading